Evaluating host range in weed biocontrol agents: insights from a comparison of native range natural enemy surveys and laboratory host tests

Introducing exotic insects for classical biological control of weeds carries risks, necessitating rigorous testing to ensure candidate agents pose little threat to non-target species. However, host testing methods can overestimate the likelihood of damage to non-target plants under field conditions. Consequently, effective, host-specific biocontrol agents can be unnecessarily denied permission for release. In assessing the host range of two weevil species as candidate biocontrol agents for the aquatic weed, Sagittaria platyphylla (Engelm.) J.G.Sm. (Alismataceae) we designed native range surveys and laboratory host tests, focusing on closely related species present in both contexts. This allowed us to relate host testing results to field host use of a fruit-feeding weevil, Listronotus appendiculatus LeConte (Coleoptera: Curculionidae) and a crown-boring weevil, L. sordidus Gyllenhal. We used molecular barcoding of endogenous larvae to identify distinct Listronotus taxa associated with each sampled host plant species in the North American native range. In laboratory tests, both Listronotus species could oviposit and develop to adults on several Alismataceae genera in addition to Sagittaria. Field surveys revealed a clear distinction between the Listronotus taxa associated with Sagittaria and those sampled from Echinodorus, indicating that L. appendiculatus and L. sordidus don’t utilise Echinodorus in the field. Field surveys found that laboratory host testing overestimated the host range of these candidate biocontrol agents. We used this information to recommend experimental approaches to better assess the host range of the candidate biocontrol agents.